How to formulate hypotheses?
Table of content
What is a hypothesis?
A hypothesis states your predictions about what your research will find. It is a tentative answer to your research question that has not yet been tested. More specifically, a hypothesis is a statement that can be tested by scientific research. If you want to test a relationship between two or more things, you need to write hypotheses before you start your experiment or data collection.
Example: Viewers exposed to a YouTube product review video with sponsorship disclosure will exhibit higher perceived persuasive intent than will those exposed to a product review video without sponsorship disclosure.
A hypothesis is not just a guess. This means that the hypothesis should be built on existing theories and previous research. A hypothesis also needs to be statistically testable, which means you can support or refute it through a scientific research methodology (e.g., experiments, surveys) and statistical analysis of data (e.g., t-test, Chi-square, ANOVA, linear regression).
The role of variables in hypotheses development
Hypotheses propose a relationship between two or more variables. An independent variable is something the researcher changes or controls (for instance, with an experiment). A dependent variable is something the researcher observes and measures (for instance, a likert scale measuring brand satisfaction).
Example: Viewers exposed to a YouTube product review video with sponsorship disclosure will exhibit less favorable attitudes toward (a) the reviewed product, (b) the reviewed brand, and (c) the reviewer, than will those exposed to a product review video without sponsorship disclosure.
In this example, the independent variable is sponsorship disclosure — the assumed cause. The dependent variables are (1) attitude toward the reviewed product, (2) attitude toward reviewed brand, and (3) attitude toward the reviewer.
Hypotheses formulation
There are many ways of formulating hypotheses. Formulating a hypothesis dependant on the type of variables included in the relationship and the appropritate statistical test needed to test the hypothesis. The following table gives some examples that may guide you in this process.
| Hypothesis formulation | Independent variable (Type of variable) | Dependent variable (Type of variable) | Moderator variable (Type of variable) | Statistical test |
|---|---|---|---|---|
| Personalized ads will be perceived as more perceived considerate treatment than non-personalized ads | Ad personalization (Qualitative, binary) | Perceived considerate treatment (Quantitative) | — | T-test |
| Personalized ads will induce more reactance to the advertisement than non-personalized ads | Ad personalization (Qualitative, binary) | Reactance to the advertisement (Quantitative) | — | T-test |
| Women are more loyal to the brand than men | Gender (Qualitative, binary) | Brand loyalty (Quantitative) | — | T-test |
| Reactance to advertisement is negatively influenced by perceived considerate treatment | Reactance to the advertisement (Quantitative) | Perceived considerate treatment (Quantitative) | — | Linear regression |
| Self-brand connection positively influences brand satisfaction | Self-brand connection (Quantitative) | Brand satisfaction (Quantitative) | — | Linear regression |
| The more brand satisfaction the more brand loyalty | Brand satisfaction (Quantitative) | Brand loyalty | — | Linear regression |
| Perceived brand heritage has a positive influence on brand trust | Perceived brand heritage (Quantitative) | Brand trust (Quantitative) | — | Linear regression |
| Organizational attractiveness exerts a positive influence on job-pursuit intention and click intention | Organizational attractiveness (Quantitative) | Job-pursuit intention (Quantitative) Click intention (Quantitative) | — | Multiple linear regression |
| Perceived intrusiveness negatively influences the attitude toward the ad, click intention, and job‐pursuit intention | Perceived intrusiveness (Quantitative) | Attitude toward the ad (Quantitative) Click intention (Quantitative) Job‐pursuit intention (Quantitative) | — | Multiple linear regression |
| There is an interaction effect between ad personalization and ad targeting on the attitude toward the ad so that personalization has positive effects when the ad is targeted and negative effects when it is not targeted | Ad personalization (Qualitative, binary) | Attitude toward the ad (Quantitative) | Ad targeting (Qualitative, binary) | Two-way ANOVA Conditional process analysis |
| The positive effects of ad personalization on perceived considerate treatment are greater for individuals with a stronger sense of uniqueness than individuals with a weaker sense of uniqueness | Ad personalization (Qualitative, binary) | Perceived considerate treatment (Quantitative) | Sense of uniqueness (Quantitative) | Conditional process analysis |
| The negative effects of ad personalization on reactance to the advertisement are lower for individuals with a stronger sense of uniqueness than individuals with a weaker sense of uniqueness | Ad personalization (Qualitative, binary) | Reactance to the advertisement (Quantitative) | Sense of uniqueness (Quantitative) | Conditional process analysis |
Note: It is also possible to formulate mediation hypotheses but this is not developed in this article.