Scribd
Upload
350 views

7 Steps of Hypothesis Testing

1) The 7 steps of hypothesis testing are: specify the null and alternative hypotheses, choose a significance level, select the appropriate test and calculate the test statistic, state the decision rule, calculate the test statistic from the sample data, use it to make a decision, and interpret the decision in context. 2) The example tests if the mean sodium content of 274 mg/kg is significantly higher than the certified value of 250 mg/kg using a two-tailed t-test at α=0.05. 3) The t-statistic of 3.024 exceeds the critical value of 2.447, so the null hypothesis that the mean is 250 mg/kg is rejected. However, the results require

Uploaded by

Shaira Mae Lapaz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
350 views

7 Steps of Hypothesis Testing

1) The 7 steps of hypothesis testing are: specify the null and alternative hypotheses, choose a significance level, select the appropriate test and calculate the test statistic, state the decision rule, calculate the test statistic from the sample data, use it to make a decision, and interpret the decision in context. 2) The example tests if the mean sodium content of 274 mg/kg is significantly higher than the certified value of 250 mg/kg using a two-tailed t-test at α=0.05. 3) The t-statistic of 3.024 exceeds the critical value of 2.447, so the null hypothesis that the mean is 250 mg/kg is rejected. However, the results require

Uploaded by

Shaira Mae Lapaz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Seven steps of hypothesis testing

Let us perform hypothesis testing through the following 7 steps of the procedure:
Step 1 : Specify the null hypothesis and the alternative hypothesis
Step 2 : What level of significance?
Step 3 : Which test and test statistic to be performed?
Step 4 : State the decision rule
Step 5 : Use the sample data to calculate the test statistic
Step 6 : Use the test statistic result to make a decision
Step 7 : Interpret the decision in the context of the original question
To guide us through the steps, let us use the following example.
Assume a food laboratory analyzed a certified reference freeze-dried food material with
a stated sodium (Na) content of 250 mg/kg. It carried out 7 repeated analyses and
obtained a mean value of 274 mg/kg of sodium with a sample standard deviation of 21
mg/kg. Now we want to know if the mean value of 274 mg/kg is significantly larger than
the stated amount of 250 mg/kg. If so, we will conclude that the reported results of this
batch of analysis were of bias and had consistently given higher values than expected.

Step 1 : Specify the null hypothesis and the alternative hypothesis


The null hypothesis Ho is the statement that we are interested in testing. In this case,
the null condition is that the mean value is 250 mg/kg of sodium.
The alternative hypothesis H1 is the statement that we accept if our sample outcome
leads us to reject the null hypothesis. In our case, the alternative hypothesis is that the
mean value is not equal to 250 mg/kg of sodium. In other words, it can be significantly
larger or smaller than the value of 250 mg/kg.
So, our formal statement of the hypotheses for this example is as follows:

Ho : 𝑥̅ = 250 mg/kg (i.e., the certified value)

H1 : 𝑥̅ ≠ 250 mg/kg (i.e., indicating that the laboratory has a bias result)

1
Step 2 : What level of significance
The level of significance is the probability of rejecting the null hypothesis by chance
alone. This could happen from sub-sampling error, methodology, analyst’s technical
competence, instrument drift, etc. So, we have to decide on the level of significance to
reject the null hypothesis if the sample result was unlikely given the null hypothesis was
true.

Traditionally, we define the unlikely (given by symbol ) as 0.05 (5%) or less. However,
there is nothing to stop you from using  = 0.1 (10%) or  = 0.01 (1%) with your own
justification or reasoning.
In fact, the significance level sometimes is referred to as the probability of a Type I
error. A Type I error occurs when you falsely reject the null hypothesis on the basis of
the above-mentioned errors. A Type II error occurs when you fail to reject the null
hypothesis when it is false.

Step 3 : Which test and test statistic?


The test statistic is the value calculated from the sample to determine whether to reject
the null hypothesis. In this case, we use Student’s t-test statistic in the following
manner:
𝑠
𝜇 = 𝑥̅ ± 𝑡(𝛼=0.05,𝑣=𝑛−1)
√𝑛

|𝑥̅ −𝜇|√𝑛
or 𝑡(𝛼=0.05,𝑣=𝑛−1) =
𝑠
By calculation, we get a t-value of 3.024 at the significance level of  = 0.05 and v = (7-
1) or 6 degrees of freedom for n = 7 replicates.

Step 4 : State the decision rule


The decision rule is always of the following form:
Reject Ho if …..
We reject the null hypothesis if the test statistic is larger than a critical value
corresponding to the significance level in step 2.

2
There is now a question in H1 on either one-tailed (> or <) or two-tailed (≠ not equal)
tests to be addressed. If we are talking about either “greater than” or “smaller than”, we
take the significance level at  = 0.05 whilst for the unequal (that means the result can
be either larger or smaller than the certified value), the significance level at  = 0.025 on
either side of the normal curve is to be studied.
As our H1 is for the mean value to be larger or smaller than the certified value, we use
the 2-tailed t-test for  = 0.05 with 6 degrees of freedom. In this case, the t-critical value
at  = 0.05 and 6 degrees of freedom is 2.447 from the Student’s t-table or from using
the Excel function “=T.INV.2T(0.05,6)” or “=TINV(0.05,6) in older Excel version.
That means the decision rule would be stated as below:
Reject Ho if t > 2.447

Step 5 : Use the sample data to calculate the test statistic


Upon calculation on the sample data, we have got a t-value of 3.024 at the significance
level of  = 0.05 and v = (7-1) or 6 degrees of freedom for n = 7 replicates.

Step 6 : Use the test statistic to make a decision


When we compare the result of step 5 to the decision rule in step 4, it is obvious that
3.024 is greater than the t-critical value of 2.447, and so we reject the null hypothesis. In
other words, the mean value of 274 mg/kg is significantly different from the certified
value of 250 mg/kg. Is it really so? We must go to step 7.

Step 7 : Interpret the decision in the context of the original question


Since hypothesis testing involves some kind of probability under the disguise of
significance level, we must interpret the final decision with caution. To say that a result
is “statistically significant” sounds remarkable, but all it really means is that it is more
than by chance alone.
To do justice, it would be useful to look at the actual data to see if there are one or more
high outliers pulling up the mean value. Perhaps increasing the number of replicates
might show up any undesirable data. Furthermore, we might have to take a closer look
at the test procedure and the technical competence of the analyst to see if there were
any lapses in the analytical process. A repeated series of experiment should be able to
confirm these findings.

You might also like