在R中针对实验多选择题的字符串型作答数据实现自动评分的方法

Here's a practical, step-by-step implementation to calculate category-based total scores using keyword matching (like option letters) for your experiment data:

Step 1: Load and Prepare Data #

First, we’ll read your CSV file and separate the answer key (first row) from participant responses.

# Replace with your actual file path
experiment_data <- read.csv("your_experiment_data.csv", stringsAsFactors = FALSE)

# Extract the answer key (first row holds correct answers)
answer_key <- experiment_data[1, ]

# Isolate participant responses (all rows after the first)
participant_responses <- experiment_data[-1, ]

Step 2: Define Category Prefixes #

List the category prefixes you need to score:

category_prefixes <- c("PRE_TR", "PRE_IS", "PRE_RULE", "POST_TR", "POST_IS", "POST_RULE")

Step 3: Create a Scoring Helper Function #

This function calculates total correct responses for a single category and participant. It extracts the option letter (e.g., A, B, C) from the correct answer and checks if the participant’s response contains that keyword.

# Load stringr for robust keyword extraction (optional but recommended)
library(stringr)

calculate_category_score <- function(participant_row, answer_key, category_prefix) {
  # Filter columns belonging to the current category
  category_cols <- grepl(paste0("^", category_prefix), names(participant_row))
  
  # Get participant's answers and correct answers for these columns
  participant_answers <- participant_row[category_cols]
  correct_answers <- answer_key[category_cols]
  
  # Extract the option letter (works for formats like "A. Correct Option" or "(B) Another Option")
  correct_keywords <- str_extract(correct_answers, "[A-Z]")
  
  # Check if each response contains the correct keyword
  # Set ignore.case = TRUE if responses might use lowercase letters
  is_correct <- mapply(function(resp, key) {
    !is.na(resp) && grepl(key, resp, ignore.case = FALSE)
  }, participant_answers, correct_keywords)
  
  # Sum the number of correct responses
  sum(is_correct)
}

Step 4: Calculate Scores for All Participants and Categories #

Apply the helper function to generate total scores for each category:

# Calculate category totals for all participants
category_scores <- as.data.frame(sapply(category_prefixes, function(prefix) {
  apply(participant_responses, 1, calculate_category_score, answer_key = answer_key, category_prefix = prefix)
}))

# Rename columns to match your desired output (e.g., TOTAL_PRE_TR)
colnames(category_scores) <- paste0("TOTAL_", colnames(category_scores))

Step 5: Combine Data and Export #

Merge the original participant data with the new total scores and save the result:

# Combine responses and scores
final_scored_data <- cbind(participant_responses, category_scores)

# Export to CSV (optional)
write.csv(final_scored_data, "scored_experiment_data.csv", row.names = FALSE)

Key Notes & Adjustments #

• Keyword Extraction: If your correct answers use a different format (e.g., lowercase letters, numbers), modify the str_extract pattern (e.g., [a-z] for lowercase, [0-9] for numbers).
• NA Handling: The function treats unanswered questions (NA) as incorrect. If you want to exclude them from scoring, add na.rm = TRUE to the sum() call.
• Case Sensitivity: Change ignore.case = FALSE to TRUE if participant responses might use lowercase letters for options.

Example Output Structure #

Your final data will include all original participant columns plus new columns like:

内容的提问来源于stack exchange，提问作者gokgok