Counted, Controlled or Ignored? Age as a Variable in Marketing Meta-Analyses: An Umbrella Review
Author
Kloknerová, Chiara
Term
4. Semester
Publication year
2026
Submitted on
2026-09-15
Abstract
This thesis investigates how age is treated in meta-analyses within marketing and consumer research, a field that has long argued that age is multidimensional but in practice almost always measures it chronologically. Using an umbrella review, the study conducts a manual journal-based search of eight leading outlets for meta-analyses published between 2016 and 2026, identifying 68 meta-analyses that meet the inclusion criteria. At the meta-analysis level, age is coded according to whether it is examined as a substantive construct, included only as a control variable, or not analysed at all. For the 29 meta-analyses that analyse age, every reported age effect is coded in terms of age construct, measurement form, analytical role, and the direction of the reported relationship, yielding 124 distinct effects. Because these effects estimate relationships across 57 different dependent variables, their directions are synthesised by vote counting rather than a single pooled parameter. The results show that age is variously counted, controlled, or ignored, and most often ignored. Sixteen meta-analyses treat age as a substantive variable, 13 include age as a statistical control, and 39 do not analyse age. Across all 124 effects, age is measured only as chronological age, with no alternative age constructs appearing in the database. In 95 effects, age represents the mean age of the study sample, which shifts the focus from variability between consumers to variability between studies. Around two thirds of the effects report no significant relationship, while the remaining effects split almost evenly between increases and decreases (20 increases and 22 decreases), a pattern statistically indistinguishable from chance (sign test, two-sided p = .88). Eight meta-analyses report sample mean ages between 24.9 and 33.2 years. Taken together, the evidence does not show that age is unrelated to customer outcomes, but rather that the field’s cumulative knowledge about age is built almost exclusively on one dimension – chronological age. What marketing knows about age is therefore bounded by how age has been measured, with implications for theory development, meta-analytic practice, and future research on age-related consumer behaviour.
Dette speciale undersøger, hvordan alder behandles i meta-analyser inden for marketing og forbrugerforskning, hvor alder længe er blevet forstået som et multidimensionelt fænomen, men i praksis næsten altid måles kronologisk. Gennem et såkaldt umbrella review gennemføres en manuelt baseret journalsøgning i otte førende tidsskrifter for meta-analyser publiceret mellem 2016 og 2026, hvilket resulterer i 68 meta-analyser, der opfylder inklusionskriterierne. På meta-analyseniveau kodes alder efter, om den undersøges som et substantielt begreb, kun indgår som kontrolvariabel eller slet ikke analyseres. For de 29 meta-analyser, der analyserer alder, kodes hver rapporteret alders-effekt med hensyn til alderskonstruktion, måleform, analytisk rolle og retning af den rapporterede sammenhæng, hvilket giver i alt 124 effekter. Effekternes retning syntetiseres ved hjælp af vote counting, da de dækker 57 forskellige afhængige variabler uden én fælles populationsparameter. Resultaterne viser, at alder enten tælles, kontrolleres eller ignoreres – og oftest ignoreres. Seksten meta-analyser behandler alder som en substantiel variabel, 13 anvender alder som statistisk kontrol, og 39 analyserer ikke alder overhovedet. I samtlige 124 effekter måles alder kronologisk; ingen alternative alderskonstruktioner fremkommer i databasen. I 95 effekter repræsenterer aldersvariablen gennemsnitsalderen i det enkelte studie, hvilket betyder, at variationen analyseres mellem studier snarere end mellem forbrugere. Omkring to tredjedele af effekterne viser ingen signifikant sammenhæng, mens de resterende fordeler sig næsten ligeligt på positive og negative effekter (20 stigninger og 22 fald), en fordeling der statistisk ikke kan skelnes fra tilfældighed (sign-test, tosidet p = .88). Otte meta-analyser rapporterer gennemsnitsalderen for deres stikprøver, som ligger mellem 24,9 og 33,2 år. Samlet peger resultaterne ikke på, at alder er uvedkommende for kunderesultater, men snarere på, at den eksisterende viden i marketing primært bygger på én aldersdimension, nemlig kronologisk alder. Hvad marketing ved om alder, er dermed begrænset af, hvordan alder er blevet målt, hvilket har konsekvenser for både teoriudvikling, metaanalytisk praksis og fremtidig forskning i aldersrelateret forbrugeradfærd.
[This abstract has been generated with the help of AI directly from the project full text]