# 36 subitizing trials per participant: all participants are above 90%, the worst 34 out of 36 are correct
correct_trial_list = list()
for sub_df in subitizing_df_list:
correct_trial_list.append((sub_df["deviation_score"] == 0).sum())
# remove subitizing trials
df_list_t1 = list()
for df in df_list:
df_list_t1.append(df.loc[df["numerosity"] > 4])
min_res = 10
max_res = 150
df_list_prepro = list()
for df in df_list_t1:
df_list_prepro.append(drop_df_rows_according2_one_col(df, "responseN", min_res, max_res))
# concat all participant
df_data = pd.concat(df_list_prepro)
# keep data within 3 sd
n_discs = [51, 54, 57, 60, 63, 66, 69, 72,
78, 81, 84, 87, 90, 93, 96, 99]
df_list_by_num = [get_sub_df_according2col_value(df_data, "numerosity", n) for n in n_discs]
prepro_df_list = list()
for sub_df in df_list_by_num:
lower_bondary = get_mean(sub_df, "responseN") - 3 * get_std(sub_df, "responseN")
upper_bondary = get_mean(sub_df, "responseN") + 3 * get_std(sub_df, "responseN")
new_sub_df = drop_df_rows_according2_one_col(sub_df, "responseN", lower_bondary, upper_bondary)
prepro_df_list.append(new_sub_df)
# load raw data
mydata = preprocess_exp3a_func(DATA_PATH, FILETYPE, FILENAME_PREFIX)
# preprocess starts here
mydata = keep_valid_columns(mydata, KEPT_COL_NAMES)
# drop practice trials: drop all rows with NaNs in key_resp.keys
col_to_dropna = ['key_resp.keys']
mydata = drop_df_nan_rows_according2cols(mydata, col_to_dropna)
# drop too fast and too slow response
if drop_fastandslow_resp:
col_to_drop_rows = "key_resp.rt"
min_rt = 0.15
max_rt = 3
mydata = drop_df_rows_according2_one_col(mydata, col_to_drop_rows,
min_rt, max_rt)
# add numerosity difference between D1 and D2
mydata["dff_D1D2"] = mydata["D1numerosity"] - mydata["D2numerosity"]
# add correct answer
insert_new_col_from_two_cols(mydata, "ref_first", "key_resp.keys",
"is_resp_ref_more", insert_is_resp_ref_more)
insert_new_col(mydata, "is_resp_ref_more", "is_resp_probe_more",
insert_is_resp_probe_more)
# add probe numerosity
insert_new_col_from_three_cols(mydata, "D1numerosity", "D2numerosity",
"ref_first", "probeN", insert_probeN)
# add ref numerosity
insert_new_col_from_three_cols(mydata, "D1numerosity", "D2numerosity",
"ref_first", "refN", insert_refN)
# add probe crowding condition
return df[col_name].mean()
if __name__ == '__main__':
debug = False
write_to_excel = False
# read the totalData file
all_df = pd.read_excel('../../data/exp2_data_online/clean_totalData.xlsx',
index_col=0)
# drop obvious wrong response
col_to_drop_rows = "responseN"
min_res = 10
max_res = 100
all_df = drop_df_rows_according2_one_col(all_df, col_to_drop_rows, min_res,
max_res)
# drop outside 3 strd
n_discs = [34, 36, 38, 40, 42, 44, 58, 60, 62, 64, 66, 68]
df_list = [
get_sub_df_according2col_value(all_df, "Numerosity", n)
for n in n_discs
]
col_to_process = "responseN"
prepro_df_list = list()
for numerosity, sub_df in zip(n_discs, df_list):
lower_bondary = get_mean(
sub_df, col_to_process) - 3 * get_std(sub_df, col_to_process)
upper_bondary = get_mean(