How to Calculate Conditional Relative Frequency from a Two-Way Table

Understanding the Two-Way Frequency Table

A two-way frequency table, also known as a contingency table, is a statistical tool used to display the frequencies, or counts, for two distinct categorical variables simultaneously. This structure is crucial for analyzing potential relationships or associations between these two variables. By organizing the data such that one variable defines the rows and the other defines the columns, the table allows us to immediately see how categories are distributed across both dimensions.

For illustrative purposes, consider a sample survey of 100 participants investigating preferred sports among baseball, basketball, and football, categorized by gender. The two categorical variables are Gender (Male/Female) and Favorite Sport. This matrix allows for a rapid assessment of whether sport preferences differ based on the respondent’s gender.

Defining Key Frequencies (Joint and Marginal)

Accurate calculation of conditional relative frequencies hinges on correctly identifying the two types of frequencies present in the table. The numbers residing in the interior cells of the table are known as joint frequencies. A joint frequency represents the count of observations that satisfy both the row category and the column category simultaneously—for example, the number of individuals who are Male and prefer Baseball.

In contrast, the totals positioned along the perimeter of the table (the bottom row and the rightmost column) are referred to as marginal frequencies. These totals represent the frequency of a single variable’s category without reference to the other variable. For instance, the total number of Females, regardless of their sport preference, is a marginal frequency. The overall total sample size (100 in this case) is also a marginal frequency, known as the grand total.

Interpreting the example table: The survey involved 100 people in total. The marginal totals reveal that 48 were males and 52 were females. Regarding sport preference, 36 chose baseball, 31 chose basketball, and 33 chose football. The joint frequencies provide the granular details: 13 males prefer baseball, 23 females prefer baseball, 15 males prefer basketball, 16 females prefer basketball, 20 males prefer football, and 13 females prefer football. These joint counts form the numerators for our conditional calculations.

Methodology for Conditional Relative Frequencies

The primary purpose of organizing data into a two-way frequency table is to simplify the calculation of conditional relative frequencies. These calculations are fundamentally different from general relative frequencies because they rely on a specific, predetermined condition, effectively redefining the total population being considered.

The calculation follows a simple ratio: the joint frequency (the count of the desired outcome within the condition) divided by the relevant marginal frequency (the total count defining the condition). The key challenge is correctly identifying the marginal frequency that serves as the denominator. If the condition is based on a row variable (e.g., “given they are male”), the row total is the denominator. If the condition is based on a column variable (e.g., “given they like football”), the column total is the denominator. We do not use the grand total (100) for conditional relative frequencies.

The following examples provide a step-by-step application of this methodology, demonstrating how to extract the correct numerator (joint frequency) and denominator (marginal frequency) based on the stated condition.

Example 1: Conditioning on a Row Variable (Male/Basketball)

What is the probability that a survey respondent likes basketball the most, given that the respondent is male? The explicit condition is “male,” meaning we must restrict our focus solely to the marginal total of male respondents. This total becomes the denominator.

We locate the joint frequency corresponding to the intersection of “Male” and “Basketball” and divide it by the total number of males:

• Joint Frequency (Male and Basketball): 15
• Marginal Frequency (Total Males): 48
• Calculation: 15 / 48 = 0.3125

Thus, the conditional probability that a respondent prefers basketball, given they are male, is 0.3125, or 31.25%.

Example 2: Conditioning on a Row Variable (Female/Baseball)

We are asked to find the probability that a respondent likes baseball the most, given that the respondent is female. The condition dictates that we must only consider the row corresponding to female respondents, using the total number of females as our new sample space.

We divide the joint frequency of “Female and Baseball” by the marginal frequency (total count) of “Female” respondents:

• Joint Frequency (Female and Baseball): 23
• Marginal Frequency (Total Females): 52
• Calculation: 23 / 52 = 0.4423

The conditional probability that a respondent prefers baseball, given they are female, is approximately 0.4423, or 44.23%.

Example 3: Conditioning on a Column Variable (Football/Male)

What is the probability that a survey respondent is male, given that the respondent likes football the most? In this case, the conditioning is based on the column variable (“Football”). This necessitates focusing exclusively on the Football column, and its total becomes the denominator.

We calculate this by dividing the joint frequency of “Male and Football” by the marginal frequency (total count) of all respondents who prefer “Football”:

• Joint Frequency (Male and Football): 20
• Marginal Frequency (Total Football): 33
• Calculation: 20 / 33 = 0.6060

Therefore, the probability that a respondent is male, given their preference for football, is 0.606, or 60.6%.

Example 4: Conditioning on a Column Variable (Baseball/Female)

We seek the probability that a survey respondent is female, given that the respondent likes baseball the most. The condition requires restricting the analysis to the Baseball column, which dictates the denominator.

We divide the joint frequency of “Female and Baseball” by the marginal frequency (total count) of respondents who prefer “Baseball”:

• Joint Frequency (Female and Baseball): 23
• Marginal Frequency (Total Baseball): 36
• Calculation: 23 / 36 = 0.6389

The resulting conditional probability that a respondent is female, given their preference for baseball, is approximately 0.6389, or 63.89%.

Example 5: Compound Event Condition (Male/Baseball OR Football)

What is the probability that a survey respondent likes baseball or football the most, given that the respondent is male? The condition is “Male,” fixing the denominator at 48. Since the outcome involves a compound “OR” event, we must sum the relevant joint frequencies within the specified row.

The calculation involves dividing the combined joint frequency by the marginal frequency of “Male” respondents:

• Joint Frequency (Male and Baseball OR Football): 13 + 20 = 33
• Marginal Frequency (Total Males): 48
• Calculation: 33 / 48 = 0.6875

The conditional probability that a male prefers baseball or football is 0.6875, or 68.75%.

Example 6: Compound Event Condition (Female/Baseball OR Basketball)

We are asked for the probability that a respondent likes baseball or basketball the most, given that the respondent is female. The condition is “Female,” making the denominator 52. We sum the joint frequencies for females who prefer baseball and females who prefer basketball.

We divide the combined joint frequency by the marginal frequency of “Female” respondents:

• Joint Frequency (Female and Baseball OR Basketball): 23 + 16 = 39
• Marginal Frequency (Total Females): 52
• Calculation: 39 / 52 = 0.75

The probability that a female prefers baseball or basketball, given the condition, is exactly 0.75, or 75%.

Example 7: Complementary Event Condition (Male/NOT Football)

What is the probability that a survey respondent does not like football the most, given that the respondent is male? The condition remains “Male” (Total = 48). The event “does not like football” is the complement of “likes football,” which in this context is equivalent to liking baseball or basketball.

We calculate this by summing the joint frequencies for males who prefer Baseball or Basketball, and dividing by the Male marginal total:

• Joint Frequency (Male and NOT Football, i.e., Baseball + Basketball): 13 + 15 = 28
• Marginal Frequency (Total Males): 48
• Calculation: 28 / 48 = 0.5833

The conditional probability that a male respondent does not prefer football is approximately 0.5833, or 58.33%. This final example illustrates that complex conditional statements can often be resolved by identifying the equivalent sum of complementary joint frequencies within the conditioned group.