PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 9762873

Using binary logistic regression models for ordinal data with non-proportional odds.

R Bender·U Grouven
View original source ↗Export metadata ↓

Abstract

The proportional odds model (POM) is the most popular logistic regression model for analyzing ordinal response variables. However, violation of the main model assumption can lead to invalid results. This is demonstrated by application of this method to data of a study investigating the effect of smoking on diabetic retinopathy. Since the proportional odds assumption is not fulfilled, separate binary logistic regression models are used for dichotomized response variables based upon cumulative probabilities. This approach is compared with polytomous logistic regression and the partial proportional odds model. The separate binary logistic regression approach is slightly less efficient than a joint model for the ordinal response. However, model building, investigating goodness-of-fit, and interpretation of the results is much easier for binary responses. The careful application of separate binary logistic regressions represents a simple and adequate tool to analyze ordinal data with non-proportional odds.

Explore related subjects

BiasBlood PressureDiabetes Mellitus, Type 1Diabetic RetinopathyFollow-Up StudiesGlycated HemoglobinHumansLogistic ModelsOdds RatioReproducibility of ResultsRisk FactorsSmokingComparative StudyJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

R Bender, U Grouven. 1998. Using binary logistic regression models for ordinal data with non-proportional odds.. https://doi.org/10.1016/s0895-4356(98)00066-3

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Assessment of blinding in pharmacotherapy and noninvasive neuromodulation randomized controlled trials for neuropathic pain in adults.

In randomized controlled trials (RCTs), study participants and research personnel are often blinded to minimize biases related to knowing treatment allocation. To determine if blinding was effective, participants may be asked which treatment they believe they received ("treatment guess"). This descriptive review characterized blinding assessment (BA) reporting in pharmacotherapy and neuromodulation neuropathic pain RCTs. Of 288 papers, 36 (12.5%) reported a BA. One paper reported the results of 2 studies, so in total 37 studies with a BA were assessed. Of these, 19 were crossover, 17 parallel, and 1 partial crossover in design. All 37 studies assessed participant blinding, and 10 also assessed investigator blinding. Approximately 27% included an "unsure" answer option for treatment guess, and 38% asked the reason for the guess. There were no clear patterns in BA reporting across time nor based on treatment type. Seventeen trials provided sufficient data to calculate Bang Blinding Index (BI) to determine blinding success. Participants remained blinded (BI = 0 &#xb1; 0.2) in 10/17 placebo and 10/17 treatment arms, 6 placebo and 5 treatment arms had a BI > 0.2 suggesting possible unblinding, whereas 1 placebo and 2 treatment arms had a BI < -0.2 suggesting misinformed guessing. Overall, we found that BAs are done in a minority of published neuropathic pain trials and with variable methodology. Given the importance of minimizing risk of bias because of treatment unblinding, future studies should consider including BAs, and further consensus building is necessary to determine if and how BAs should be conducted and interpreted in analgesic clinical trials.

Bias↗
PubMed2000

Gigerenzer's normative critique of Kahneman and Tversky.

Gigerenzer has argued that it may be inappropriate to characterize some of the biases identified by Kahneman and Tversky as "errors" or "fallacies," for three reasons: (a) according to frequentists, no norms are appropriate for single-case judgments because single-case probabilities are meaningless; (b) even if single-case probabilities make sense, they need not be governed by statistical norms because such norms are "content-blind" and can conflict with conversational norms; (c) conflicting statistical norms exist. I try to clear up certain misunderstandings that may have hindered progress in this debate. Gigerenzer's main point turns out to be far less extreme than the position of "normative agnosticism" attributed to him by Kahneman and Tversky: Gigerenzer is not denying that norms appropriate for single-case judgments exist, but is rather complaining that the existence and the nature of such norms have been dogmatically assumed by the heuristics and biases literature. In response to this complaint I argue that single-case probabilities (a) make sense and (b) are governed by probabilistic norms, and that (c) the existence of conflicting statistical norms may be less widespread and less damaging than Gigerenzer thinks.

Bias↗
PubMed2000

First steps in analysing NHS waiting times: avoiding the 'stationary and closed population' fallacy.

The aim of this paper is to demonstrate the effect of excluding incomplete observations and competing events when calculating cross-sectional measures of NHS waiting times, and to obtain a more accurate estimate of the 'time-to-admission' of those listed on NHS waiting lists using life-table methods. The official 'times-since-enrollment' of all elective 'admissions' in England, 1 July to 31 December 1994 inclusive, were extracted from Hospital Episode Statistics. The official 'times-to-census' of all those on a waiting list in England at 30 September 1994 were obtained from aggregated KH07 data. The percentage waiting at least three months, at least six months etc., was calculated separately for each data set and compared with a period life-table derived from the combined data. The cumulative likelihood of elective admission is markedly overestimated across the whole range of waiting times. The experience of those still waiting, those removed from the list, those suspended or deferred and those put to the back of the queue is not taken into account in the calculation of official waiting times. The Department of Health currently presents the 'time-since-enrollment' of those admitted as though it indicates how long all patients can expect to wait for admission. The consequent bias in published summary statistics incorrectly quantifies the real experience of patients. It is recommended that calculation of waiting times from KH07 census counts and Hospital Episode Statistics be reconsidered in the light of what patients, clinicians, managers and politicians need to know about treatment delay.

Bias↗