Variables, Research Questions & Hypotheses

• Variables:

  • an element, feature, or factor that is liable to vary or change
  • any factor that can be controlled, changed, or measured
  • In experimental designs we classify variables as dependent, independent or co-variables
• Independent variable:

  • Is the variable that is hypothesised to influence another variable
  • Can be a condition, intervention or a characteristic
  • Is the variable that we often experimentally manipulate to then observe the outcome
• Dependent variable:

  • Is the outcome variable that is hypothesised to be affected by the independent variable
  • Usually identified as the outcome measure in a study
  • It is the outcome we observe for change after manipulating the independent variable
• How to develop a research question:

  1. choose a topic
  2. narrow it down
  3. pick a focus
  4. form a research question from the focus - as small and discrete
• Research Question:

  • The research question may be open ended (research hypotheses are not
  • It should be answerable and unambiguous
  • The research question informs the choice of relevant aims and likely design
• Developing a research question:

  • Criteria used to assess or develop a research question: (FINER acronym)
  • Is the study Feasible?
  • Is the problem Interesting?
  • Is the problem Novel?
  • Can the problem be studied Ethically?
  • Is the problem professionally Relevant?
• Research Questions: Examples:

  • Does a home based leisure education programme influence wellbeing, depressive symptoms and quality of life after stroke compared to usual care? (Desrosiers et al, 2007)
  • What are the prevalence, severity and correlates of chronic pain among men with spinal cord injury? (Rintala et al, 1998)
  • What strategies do women with chronic arthritis use to manage pain and to maintain activities? How do these change over 2 years? (Goodacre, 2006)
• Developing a research hypothesis:

  • Further refines the research question by describing in operational terms exactly what researchers think will happen in the study
  • It is a prediction
• Research Hypothesis (H1)

  • Always a positive statement:
  • “CVA patients who receive early intervention will demonstrate significant improvements in upper limb function over the short and medium term compared to CVA patients for whom intervention was late or delayed"
• Null hypothesis (H0)

  • States that the relationship predicted by the research hypothesis does not exist. (not the reverse or opposite of research hypothesis)
  • So if we “find” a relationship then we reject the null hypothesis
  • Usually achieved by statistical decision-making (inference)
• Null hypothesis (H0)

  • If research is about assessing effect of a treatment, the null hypothesis:
  • “There will be no difference between groups (or treatments.)”
  • If null hypothesis cannot be rejected, then the results may be due to chance factors
• Research Null Hypothesis:

  • “CVA patients who receive early intervention will demonstrate no significant improvement in upper limb function over the short and medium term compared to CVA patients for whom intervention was late or delayed."
• One-tailed/two-tailed hypotheses:

  • Predicts the direction of the hypothesis
  • One tailed hypotheses
  • E.G. Predicts that group A, (treatment A) will be better than group B
  • Two tailed hypotheses
  • E.G predicts difference between two groups (in any direction)