t検定は,2つのグループの平均値に「統計的に意味のある差(有意差)」があるかどうかを判定するための統計手法.得られたデータに平均値の差があっても,それが「たまたま (偶然) 」 生じたものかを見極める.
具体的には,2郡の分散が大きく異ならない場合,平均・分散→統合 (プールした) 分散→t値→p値を求め,例えば有意水準5%の場合,p値 (有意確率) が基準の0.05より大きければ,「意味のある差がない」 と判断する.
Example Question (Oxford AI SL P408 Example7)
Mr Arthur gives his two chemistry groups the same test. He wants to find out whether there is any difference between the achievement levels of the two groups.
The results are:
a) Write down the null and alternative hypotheses,
a) の解答は下の"Oxford TextのSolution"の通り
b) Find the t-value and p-value for a t-test at the 5% significance level,
c) Write down the conclusion to the lest.
Excelでt値をいきなり求めるには 「データ分析」 を用いる方法しかない.
Excelでp値をいきなり求めるには,T.TEST(配列1,配列2,2,2)とすればp=0.816229568が得られる.
Oxford Textbook の Solution
a) $H_0$ : $\overline{x_1}= \overline{x_2}$ (there is no differencebetween the grades in Group 1 and the grades in Group 2)
$H_1$ : $\overline{x_1} \neq \overline{x_2}$ (there is a difference between the grades in Group 1 and the grades in Group 2)
b) t-value =-0.235, p-value = 0.816
c) 0.816 > 0.05, so you accept the null hypothesis: there is no significant difference between the two groups.
Notice that the two groups do not need to be the same size. This will be a two-tailed test as you want to know whether Group 1 is better or worse than Group 2.
****
(Oxford AI HL P603 Q12)
Lisa measures the arm spans (the length from finger tips on one hand to finger tips on the other hand, when arms are held horizontally outstretched) of students in two mathematics classes. She is interested to see whether the mean arm span of class A is the same as the mean arm span of class B. Lisa's data is shown in the table below.
In this question, you will use a t-test to compare the means of the two groups at the 10% level of significance. You may assume the data is normally distributed and the standard deviations are equal between the two groups, (等分散を仮定してよいと指示している)
a) (i) State the null hypothesis.
(ii) State the alternative hypothesis,
a) の解答は下の"Oxford TextのSolution"の通り
b) Find the associated p-value for this test,
c) State, giving a reason, whether Lisa should accept the null
hypothesis.
Excelの 「データ分析」 の結果は以下の通り.
Excelでp値をいきなり求めるには,T.TEST(配列A,配列B,2,2)とすればp=0.217565256が得られる.
Oxford Textbook の Solution
a) (i) $H_0$ : $\overline{x_A}= \overline{x_B}$
(ii) $H_0$ : $\overline{x_A} \neq \overline{x_B}$
b) p = 0.209 (3 sf) (不等分散を仮定した解答になっている)
c) As 0.209 > 0.1 there is no evidence to reject the null hypothesis at the 10% level of significance.
