The study adopted cross-sectional and community-based design, through which associations between anthropometric status and child eating behaviour among preschool children 2–4 years were determined. Consent letter was obtained from the study participants’ mothers/caregivers and fathers of the preschool children, as well as the traditional ruler of the community.

The study population was two hundred and ninety (290) preschool children aged 2–4 years living in communities in Umuahia South LGA of Abia State. The sample size for the study was estimated using Cochran’s formula elucidated by Araoye [10].

The study aimed to investigate child eating behaviour among preschool children because this is the stage where eating behaviours are formed that would trickle into adulthood thereby influence their nutritional/anthropometric status. There is urgent need to curb malnutrition menace in all its forms, and hence the study, which promises to provide information that will guide partners, implementers, governments at all level, academia, civil society and private sectors to drive policy implementation to mitigate malnutrition amongst under-five children.

$$n\, = \,\fracP(100 - P)}}}}$$

Margin of tolerable sampling error applied is 5% and 95% confidence level of the standard normal distribution curve which is Z = 1.96 will be used.

P = Prevalence of underweight among under 5 children in Nigeria which is 22% according to NDHS [2].

$$\begin} = \frac^2}X22(100 - 7)}}}} = \frac \right)}}} = \frac}} = 263.687 }\\ \end$$

To make up for dropout during data collection, 10% of the sample size was computed and added to the sample size. 263.687 × 10% of 263.687 = 263.687 + 26.3687 290.056 = 290.

Therefore, the sample size was two hundred and ninety (290) preschool children.

Prior to data collection, a non-probability purposive sampling technique was used to identify two hundred and ninety (290) households with preschool children aged 2–4 years across the five [5] communities in Umuahia South LGA. However, 256 households consented to participate in the study, and this gave 88.3% response rate (that is 256/290 × 100/1). One [1] child per household was recruited to be part of the study. In situations where a household had more than a child within the study age group, a simple random sampling by balloting without replacement was used. The ballot papers had only one “Yes” with several “No” and the children in such households were allowed to pick and the child that picked “Yes” was part of the study. However, this method was not applicable in households with only a child within the age bracket (2–4 years) for the study, and as such the child automatically becomes part of the study.

A pretested and validated semi-structured questionnaire was used to collect information on the date of birth of the preschool children, child eating behaviour and socioeconomic characteristics of their caregivers. For weight measurement, a Camry digital bathroom weighing scale was used to measure the children’s weight. The measurement was done with each child naked and barefooted, standing on the scale with head pointing straight. The weight was taken to the nearest 0.1 kg. The height measurement was carried out with calibrated height-meter (standio-meter), whereby each child was measured barefooted and the head, back, buttocks, calves and heels rested on the height rule for accurate reading of the height, while looking straight. Height measurement was recorded to the nearest 0.1 cm. Child-eating behaviour was assessed using the Child Eating Behaviour Questionnaire (CEBQ), which has eight items and contained thirty-five questions. The 8 CEBQ items include, food responsiveness (4 questions), emotional over-eating (4 questions), enjoyment of food (4 questions), desire to drink (3 questions), satiety responsiveness (5 questions), slowness in eating (4 questions), emotional undereating (4 questions) and food fussiness (7 questions).

The World Health Organization (WHO) Anthro-Software for Child Growth Standards was used to categorize the anthropometric status of the preschool children into weight-for-age, height-for-age and weight-for-height z-scores ± 2 standard deviation (SD), while child eating behaviour was assessed using a 5-point Likert scale (Never = 1; Rarely = 2; Sometimes = 3; Often = 4; and Always = 5). The mean of the responses was computed and graded as poor and good eating behaviours. Mean of 1.00–2.99 = poor eating behaviour and while mean of 3.00 and above = good eating behaviour.

Descriptive statistics– frequency and percentages was used to analyze the socioeconomic characteristics of their caregivers, and cross-tabulation was used to analyze the anthropometric status of the preschool children. Pair-sample t-test was used to determine differences in their eating behaviour by gender, while regression and correlation analysis was performed to determine the extent to which child-eating behaviour predicted anthropometric status, and also, regression analysis was performed to determine the extent to which socioeconomic status predicted eating behaviour at 5% level of significance. Data were presented in mean and standard deviation.