generalized linear models (glms) assuming binomial distributions and using the logit link function Search Results


logit binomial glm  (MathWorks Inc)
90
MathWorks Inc logit binomial glm
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Logit Binomial Glm, supplied by MathWorks Inc, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/logit binomial glm/product/MathWorks Inc
Average 90 stars, based on 1 article reviews
logit binomial glm - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
glm  (SAS institute)
90
SAS institute glm
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Glm, supplied by SAS institute, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/glm/product/SAS institute
Average 90 stars, based on 1 article reviews
glm - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
glmm  (STATA Corporation)
90
STATA Corporation glmm
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Glmm, supplied by STATA Corporation, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/glmm/product/STATA Corporation
Average 90 stars, based on 1 article reviews
glmm - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
proc logistic/ proc genmod version 9.3.1  (SAS institute)
90
SAS institute proc logistic/ proc genmod version 9.3.1
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Proc Logistic/ Proc Genmod Version 9.3.1, supplied by SAS institute, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/proc logistic/ proc genmod version 9.3.1/product/SAS institute
Average 90 stars, based on 1 article reviews
proc logistic/ proc genmod version 9.3.1 - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
proc logistic  (SAS institute)
90
SAS institute proc logistic
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Proc Logistic, supplied by SAS institute, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/proc logistic/product/SAS institute
Average 90 stars, based on 1 article reviews
proc logistic - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
proc glimmix  (SAS institute)
90
SAS institute proc glimmix
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Proc Glimmix, supplied by SAS institute, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/proc glimmix/product/SAS institute
Average 90 stars, based on 1 article reviews
proc glimmix - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
logistic general linear model (glm) matlab 2018a version 9.4.0  (MathWorks Inc)
90
MathWorks Inc logistic general linear model (glm) matlab 2018a version 9.4.0
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Logistic General Linear Model (Glm) Matlab 2018a Version 9.4.0, supplied by MathWorks Inc, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/logistic general linear model (glm) matlab 2018a version 9.4.0/product/MathWorks Inc
Average 90 stars, based on 1 article reviews
logistic general linear model (glm) matlab 2018a version 9.4.0 - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
rstudio version 1.3.1103  (RStudio)
90
RStudio rstudio version 1.3.1103
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Rstudio Version 1.3.1103, supplied by RStudio, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/rstudio version 1.3.1103/product/RStudio
Average 90 stars, based on 1 article reviews
rstudio version 1.3.1103 - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
generalized linear model (glm) function  (RStudio)
90
RStudio generalized linear model (glm) function
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Generalized Linear Model (Glm) Function, supplied by RStudio, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/generalized linear model (glm) function/product/RStudio
Average 90 stars, based on 1 article reviews
generalized linear model (glm) function - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
glm function  (CH Instruments)
90
CH Instruments glm function
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Glm Function, supplied by CH Instruments, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/glm function/product/CH Instruments
Average 90 stars, based on 1 article reviews
glm function - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
output code glm  (RStudio)
90
RStudio output code glm
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Output Code Glm, supplied by RStudio, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/output code glm/product/RStudio
Average 90 stars, based on 1 article reviews
output code glm - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier
proc corr  (SAS institute)
90
SAS institute proc corr
( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of <t>generalised</t> <t>linear</t> <t>model</t> <t>(GLM)</t> coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).
Proc Corr, supplied by SAS institute, used in various techniques. Bioz Stars score: 90/100, based on 1 PubMed citations. ZERO BIAS - scores, article reviews, protocol conditions and more
https://www.bioz.com/result/proc corr/product/SAS institute
Average 90 stars, based on 1 article reviews
proc corr - by Bioz Stars, 2026-03
90/100 stars
  Buy from Supplier

Image Search Results


( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of generalised linear model (GLM) coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).

Journal: eLife

Article Title: Clarifying the role of an unavailable distractor in human multiattribute choice

doi: 10.7554/eLife.83316

Figure Lengend Snippet: ( a ) Multiattribute stimuli in a previous study . Participants made a speeded choice between two available options (HV: high expected value; LV: low expected value) in the presence (ternary) or absence (binary) of a third distractor option (‘ D ’). Three example stimuli are labelled for illustration purposes only. In the experiment, D was surrounded by a purple square to show that it should not be chosen. ( b, c ) Relative choice accuracy (probability of H choice among all H and L choices) in ternary trials (panel b) and in binary-choice baselines (panel c) plotted as a function of both the expected value (EV) difference between the two available options (HV − LV) ( y -axis) and the EV difference between D and H ( x -axis). Relative choice accuracy increases when HV − LV increases (bottom to top) as well as when DV − HV increases (left to right, i.e., positive D effect). ( d–f ) Predicting relative accuracy in human data using regression models (Methods). Asterisks: significant effects [p < 0.05; two-sided one-sample t- tests of generalised linear model (GLM) coefficients against 0] following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant. ( g ) Rival hypotheses underlying the positive notional distractor effect on binary-choice accuracy. Left: EV indifference contour map; Right: additive utility (AU) indifference contour map. Utility remains constant across all points on the same indifference curve and increases across different curves in evenly spaced steps along the direction of the dashed grey line. Decision accuracy scales with Δ(utility) between H and L in binary choices. Left (hypothesis 1): Δ(EV) = (HV − LV)/(HV + LV) by virtue of divisive normalisation (DN); because HV 2 − LV 2 = HV 1 − LV 1 , and HV 2 + LV 2 > HV 1 + LV 1 , Δ(EV 2 ) becomes smaller than Δ(EV 1 ). Right (hypothesis 2): Δ(AU) = AU H − AU L ; Δ(AU 2 ) < Δ(AU 1 ) by virtue of additive integration. ( h, i ) Regression analysis of model-predicted accuracy in binary choice. EV + DN model corresponds to hypothesis 1 whilst AU model corresponds to hypothesis 2 in panel g. Error bars = ± standard error of the mean (SEM) ( N = 144 participants).

Article Snippet: We used logit binomial GLM to analyse trial-by-trial relative choice accuracy (Matlab’s glmfit with ‘binomial’ distribution and ‘logit’ link function).

Techniques:

The generalised linear models (GLMs) here included an additional regressor HV + LV. Error bars = ± standard error of the mean (SEM) ( N = 144 participants). Asterisks: significant effects (p < 0.05; two-sided one-sample t -tests of GLM beta coefficients against 0) following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant.

Journal: eLife

Article Title: Clarifying the role of an unavailable distractor in human multiattribute choice

doi: 10.7554/eLife.83316

Figure Lengend Snippet: The generalised linear models (GLMs) here included an additional regressor HV + LV. Error bars = ± standard error of the mean (SEM) ( N = 144 participants). Asterisks: significant effects (p < 0.05; two-sided one-sample t -tests of GLM beta coefficients against 0) following Holm’s sequential Bonferroni correction for multiple comparisons. n.s.: non-significant.

Article Snippet: We used logit binomial GLM to analyse trial-by-trial relative choice accuracy (Matlab’s glmfit with ‘binomial’ distribution and ‘logit’ link function).

Techniques: