BS1.8: Hypothesis Testing |
OBJECTIVES |
At the end of this section you should understand the concept of hypothesis testing. |
In the previous section you dealt with the first form of statistical inference, estimation. We selected a sample to draw conclusions about the population by calculating a confidence interval.
The second form of statistical inference is hypothesis testing. Hypothesis testing is a method used by researchers to determine how likely it is that observed differences between different sample estimates are entirely due to sampling error rather than underlying population differences.
Suppose a researcher wish to compare mean systolic blood pressure values between two groups (exposed and non-exposed). The following descriptive measures were computed:
Descriptive measures | Sample #1 (Exposed) |
Sample #2 (Non-exposed) |
---|---|---|
Sample size | 15 | 12 |
Mean systolic BP (mm Hg) | 130 | 120 |
Standard deviation | 22 | 20 |
We may hypothesise that the two means are similar. By means of hypothesis testing you can determine whether or not such statements are compatible with available data.
State the null hypothesis: Population means are equal (Mean1 - Mean2 = 0).
Select a level of significance: P = 0.05
The level of significance is a probability value, denoted by (P), that we use as a cut-off value by convention to reject the null hypothesis.
We reject the null hypothesis when the P-value is less than or equal to 0.05 (P £ 0.05), implying that the two means (Mean1 and Mean2) are significantly different. On the contrary, when the P-value is greater than 0.05 (P > 0.05) then we cannot reject the null hypothesis.
Also, there is a relationship between the level of significance (5%) and the level of confidence (95%), that is, 100% - 5% = level of confidence. So, if the P-value is less than 5% we are increasing level of confidence. A P-value is a measure of how much evidence we have against the null hypothesis. The smaller the P-value, the more evidence we have against the null hypothesis.